Study The Technology Of Extraction Situation Factor For Network Security Situation Awareness

With the rapid development of computer and network technology, Our lives become more convenient, However, due to the openness and sharing network, which also inevitably introduces security risks,Currently the attacks to Equipment and network services become modular, distributed, indirect, complicated than before. In this case, to carry out the research of network security situation awareness is necessary.Network security situation awareness technology is different from previous network security technology, It is a proactive network security technology, Through extraction situation factor. comprehension of current situation and projection of future status ensure network security. The extraction of situation factor is the basis of the whole process of situational awareness. This article carry out the resourch of extracting the situation factor, and do some work on the attribute reduction techniques and classification techniques involved in the process of extracting situation factor.Firstly, for using Pawlak rough set to deal with the properties of continuous type will cause the loss of information; introduce generalized rough set theory to make dimension reduction of the data. Based on the generalized rough set theory, quantificate the knowledge cotained in the attributes set and the importance of a property. and proved some related properties, on this basis an algorithm of dimension reduction is proposed.Secondly, for the BP neural network used in classification has the characteristics of easy to fall into local minimum, slow convergence.This paper designs an adaptive genetic algorithm (IGA) to optimize the parameters of BP neural networK.in the IGA the population is divided into elite groups and ordinary groups of two parts. in the course of evolution the population will adjust the intensity of mutation operator,in this way can avoid the premature of the population. Finally, experiments are carried out to analysis the performance of the proposed algorithm.through the proposed algorithm of dimension reduction operate we get a subset of the experimental data; through the experiment we know the subset of the experimental data is better than the original experimental data. Then we do experimental on the subset of data to analysis of performance of the IGA algorithm.